CN103686976A - Power control method and device in D2D (Device-to-Device) system - Google Patents

Abstract

The invention discloses a power control method and device in a D2D system, used for solving the power control problems in D2D communication. The method comprises that the position of a D2D terminal, the distance between the emission end and the reception end of a D2D subscriber and the initial D2D emission power are reported to a base station by the terminal; the base station calculates the path gain and the SINR (Signal to Interference plug Noise Ratio) of the subscriber to determinate the utility function; based on a non-cooperative game model, the base station obtains a power adjusting scheme via iterative operation; and after completion of the iterative operation, the base station sends a power control instruction to the terminal to realize power control. The method provided by the invention can decrease the calculation complexity of power control, improves the performance of the whole cellular system, and improves the utilization rate of frequency spectra.

Description

Method and implement device thereof that in a kind of D2D system, power is controlled

Technical field

The present invention relates to wireless communication field, relate in particular to the right Poewr control method of D2D user under cellular network.

Background information

Along with the broad development of the cellular radios such as LTE/LTE-Advanced, D2D(Device to Device) communicating by letter has caused more concern.

D2D communication is a kind of under the control of cellular system, between each terminal by the new technique of multiplexing cell frequency resources direct communication.Owing to itself having short haul connection feature and direct communication mode, the D2D communication technology can effectively be saved system resource, elevator system spectrum efficiency.

But, due to D2D multiplexing the frequency resource of cellular network, so just can cause interference to phone user, therefore and also can produce corresponding interference between each D2D user, introduce in D2D communication that power controls, be necessary.Power is controlled to contribute to reduce and is disturbed, and elevator system performance, reduces energy resource consumption.

At present, the power control techniques for D2D communication mainly contains the open Loop Power control technology after open Loop Power control technology conventional in LTE and improvement.These two kinds of power control techniques have larger impact for the quality of D2D communication.Given this, the present invention proposes a kind of method that power based on non-cooperative game opinion is controlled.

Summary of the invention

The invention provides a kind of power control algorithm and implement device thereof based on non-cooperative game opinion.The present invention's research is in a base station range, and a phone user and the right power control scheme of a plurality of D2D user, increase throughput of system, and elevator system performance, improves user's experience.

For solving above-mentioned technical problem, the present invention proposes following scheme, and step is as follows:

Step 1, terminal are to base station reporting information, and the information reporting comprises the position of D2D terminal, the distance of D2D terminal and base station, the Initial Trans of D2D terminal.

Calculation of parameter is carried out in step 2, base station, and calculate path gain, user's SINR base station, with this, determines utility function.

Introduce the design of utility function below.

The throughput of taking into account system and user's SINR, we have designed following utility function:

The utility function of unique user suc as formula shown in:

u _i＝B _ilog(1+γ _i)-C _ip _i

In formula, B _irepresent system bandwidth, C _ip _ithe cost function that represents increased power.Introducing cost function is because the increase meeting of D2D user power comprises that to other user phone user causes interference, γ _ithe SINR that represents D2D user i.γ _ican represent with following formula:

${\mathrm{\&gamma;}}_i=\frac{P}{P_I+{\mathrm{<figure-callout\; id="0"\; label="N"\; filenames="DEST\_PATH\_HDA0000457159260000051.png,DEST\_PATH\_HDA0000457159260000071.png"\; state="\{\{state\}\}">N</figure-callout>}}_0}=\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}$

In above formula | h _ii| ², | h _ij| ²represent path gain.

Owing to considering the effectiveness of whole system, so the design of overall utility function is shown below:

$U=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{u}_i=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(B_{i}1n(1+{\mathrm{\&gamma;}}_i)-C_i{p}_i)$

Overall utility function can represent with following formula again, wherein Γ _i(p) can be considered as the optimal compensation item to D2D user i.

$U=u_i\left(p\right)+\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)=u_i\left(p\right)+{\mathrm{\&Gamma;}}_i\left(p\right)$

First to the optimal compensation to independent variable p _iask partial derivative.

$\frac{\&PartialD;\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)}{{\&PartialD;p}_i}=\frac{\&PartialD;\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}[1n(1+{\mathrm{\&gamma;}}_j)-C_i{p}_j]}{{\&PartialD;p}_i}$

$=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{1}{1+{\mathrm{\&gamma;}}_j}\&CenterDot;\frac{{\&PartialD;\mathrm{\&gamma;}}_j}{{\&PartialD;p}_i}$

$=\underset{j\&NotEqual;\mathrm{<figure-callout\; id="1"\; label="i"\; filenames="DEST\_PATH\_HDA0000457159260000071.png"\; state="\{\{state\}\}">i</figure-callout>}}{\mathrm{\&Sigma;}}\frac{1}{1+\frac{{\left|h_{\mathrm{jj}}\right|}^2{p}_j}{\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2}}\&CenterDot;\frac{-{\left|h_{\mathrm{jj}}\right|}^2{\left|h_{\mathrm{ij}}\right|}^2{p}_j}{{[\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_{\mathrm{mj}}+{\mathrm{\&sigma;}}^2]}^2}$

$\begin{array}{c}=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{-{\left|h_{\mathrm{jj}}\right|}^2{\left|h_{\mathrm{ij}}\right|}^2{p}_j}{{[\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_{\mathrm{mj}}+{\mathrm{\&sigma;}}^2]}^2+\left[{\left|h_{\mathrm{jj}}\right|}^2{p}_j(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)\right]}\\ ={\mathrm{\&alpha;}}_i\end{array}$

In order to use the mode of iteration, in above formula, by the power division that once iteration produces, replace current power division, so above formula has just become a constant for current iteration, its integration becomes:

α _ip _i＝Γ _i(p)

It is as follows that we obtain equivalent utility function:

L＝u _i(p)+Γ _i(p)

＝ln(1+γ _i)-Cp _i+α _ip _i

The Nash Equilibrium point that can prove this utility function exists, and Power Control Problem changes the non-cooperative game model with compensation term into thus.We adopt the mode of iteration to go to approach the optimal compensation at this.

Step 3, based on non-cooperative game model, by interative computation, obtain power and adjust scheme.

To the total utility function designing in step 2 to p _iask local derviation, have:

$\frac{\&PartialD;u}{{\&PartialD;p}_i}=\frac{\&PartialD;[1n(1+{\mathrm{\&gamma;}}_i)-C_i{p}_i+\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)]}{{\&PartialD;p}_i}$

$=\frac{1}{1+{\mathrm{\&gamma;}}_i}\frac{{\&PartialD;\mathrm{\&gamma;}}_i}{{\&PartialD;p}_i}-C_i+{\mathrm{\&alpha;}}_i$

$=\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}}\frac{{\left|h_{\mathrm{ii}}\right|}^2}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}|{h_{\mathrm{ji}}|}^2{p}_j+{\mathrm{\&sigma;}}^2}-C_i+{\mathrm{\&alpha;}}_i$

Obtaining the optimized condition of this iteration is that above formula equals 0, and total utility function is to p _iask local derviation to equal 0, be shown below.

$\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}}\frac{{\left|h_{\mathrm{ii}}\right|}^2}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}+{\mathrm{\&alpha;}}_i-C_i=0$

Can obtain p thus _iiterative formula

$p_i=\frac{1}{C_i-{\mathrm{\&alpha;}}_i}-\frac{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}{{\left|h_{\mathrm{ii}}\right|}^2}$

First we set minimax transmitting power and precision parameter delta, according to iterative formula, carry out iterative computation, when this difference that calculates the utility function value that utility function value and last computation obtain is less than the precision parameter delta of setting, iteration finishes, and obtains optimal transmit power.

Power control instruction is assigned to terminal in step 4, base station, realizes power and controls.

Poewr control method based in above-mentioned D2D system, has realized as lower device, and device comprises three large modules:

System information acquisition module, is mainly used in receiving the information of reporting of user base station, upgrades D2D system phone user and D2D user's information;

Data processing module, for calculating path gain between user and user's SINR value, determines utility function;

Power regulation module, for carrying out interative computation according to the result of data processing, constantly adjusts user power and makes it reach Nash Equilibrium point, and power is adjusted to result is issued to each user.

Beneficial effect of the present invention is that the problem that power is controlled is converted into the model of non-cooperative game, has simplified computation complexity by iterative computation.

Accompanying drawing explanation

Fig. 1 is the flow chart of the inventive method

Fig. 2 is the structure chart of apparatus of the present invention

Fig. 3 is research scene graph of the present invention

Fig. 4 is cell terminal distribution schematic diagram

Fig. 5 realizes the system spectrum utilance figure that power is controlled

Fig. 6 is the availability of frequency spectrum comparison diagram of the system under different cost parameters

Fig. 7 is the availability of frequency spectrum comparison diagram of the phone user under different cost parameters

Embodiment

The present invention is a kind of method that D2D power based on non-cooperative game opinion is controlled, and concrete comprises: step 1, terminal are to base station reporting information, and the information reporting comprises the position of D2D terminal, the distance of D2D terminal and base station, the Initial Trans of D2D terminal; Calculation of parameter is carried out in step 2, base station, and calculate path gain, user's SINR base station, with this, determines utility function; Step 3, based on non-cooperative game model, by interative computation, obtain power and adjust scheme; Power control instruction is assigned to terminal in step 4, base station, realizes power and controls.

In specific implementation process, calculating and the interative computation of base station are as follows:

In step 2, the throughput of base station taking into account system and user's SINR have designed utility function, and wherein the utility function of unique user is shown below.

u _i＝B _iln(1+γ _i)-C _ip _i

In formula, B _irepresent system bandwidth, C _ip _ithe cost function that represents increased power.γ _ithe SINR that represents D2D user i.γ _ican use following formula table, wherein | h _ii| ², | h _ij| ²represent path gain.

${\mathrm{\&gamma;}}_i=\frac{P}{P_I+{\mathrm{<figure-callout\; id="0"\; label="N"\; filenames="DEST\_PATH\_HDA0000457159260000051.png,DEST\_PATH\_HDA0000457159260000071.png"\; state="\{\{state\}\}">N</figure-callout>}}_0}=\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}$

Owing to considering the effectiveness of whole system, so the design of overall utility function is shown below.

$U=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{u}_i=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(B_{i}1n(1+{\mathrm{\&gamma;}}_i)-C_i{p}_i)$

Overall utility function can represent with following formula again, wherein Γ _i(p) can be considered as the optimal compensation item to D2D user i.

$U=u_i\left(p\right)+\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)=u_i\left(p\right)+{\mathrm{\&Gamma;}}_i\left(p\right)$

First to the optimal compensation to independent variable p _iask partial derivative:

$\frac{{\&PartialD;\mathrm{\&Gamma;}}_i\left(p\right)}{{\&PartialD;p}_i}=\frac{\&PartialD;\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)}{{\&PartialD;p}_i}$

$=\frac{\&PartialD;\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}[1n(1+{\mathrm{\&gamma;}}_j)-C_j{p}_j]}{{\&PartialD;p}_i}$

$=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{1}{1+{\mathrm{\&gamma;}}_j}\&CenterDot;\frac{{\&PartialD;\mathrm{\&gamma;}}_j}{{\&PartialD;p}_i}$

$=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{1}{1+\frac{{\left|h_{\mathrm{jj}}\right|}^2{p}_j}{\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2}}\&CenterDot;\frac{-{\left|h_{\mathrm{jj}}\right|}^2{\left|h_{\mathrm{ij}}\right|}^2{p}_j}{{(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)}^2}$

$\begin{array}{c}=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{-{\left|h_{\mathrm{jj}}\right|}^2{\left|h_{\mathrm{ij}}\right|}^2{p}_j}{{(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)}^2+{\left|h_{\mathrm{jj}}\right|}^2{p}_j(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)}\\ ={\mathrm{\&alpha;}}_i\end{array}$

In order to use the mode of iteration, in above formula, by the power division that last iteration produces, replace current power division, so above formula has just become a constant for current iteration, its integration becomes:

α _ip _i＝Γ _i(p)

It is as follows that we obtain new utility function

L＝u _i(p)+Γ _i(p)

＝ln(1+γ _i)-Cp _i+α _ip _i

Total utility function under this form meets the non-cooperative game model of the optimal compensation.For realizing the optimal compensation, we adopt the method for iteration to carry out, and constantly adjust each user's transmitting power, and when utility function reaches the Nash Equilibrium point of this non-cooperative game model, total utility function reaches optimum value.

Wherein step 3 to the utility function obtaining in step 2 to p _iask local derviation, have:

$\frac{\&PartialD;u}{{\&PartialD;p}_i}=\frac{\&PartialD;[1n(1+{\mathrm{\&gamma;}}_i)-C_i{p}_i+\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)]}{{\&PartialD;p}_i}$ $=\frac{1}{1+{\mathrm{\&gamma;}}_i}\frac{{\&PartialD;\mathrm{\&gamma;}}_i}{{\&PartialD;p}_i}-C_i+{\mathrm{\&alpha;}}_i$

$=\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}}\frac{{\left|h_{\mathrm{ii}}\right|}^2}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}-C_i+{\mathrm{\&alpha;}}_i$

Obtaining the optimized condition of this iteration is that above formula equals 0, and total utility function is to p _ilocal derviation equals 0, is shown below:

$\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}}\frac{{\left|h_{\mathrm{ii}}\right|}^2}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}+{\mathrm{\&alpha;}}_i-C_i=0$

Can obtain p thus _iiterative formula

$p_i=\frac{1}{C_i-{\mathrm{\&alpha;}}_i}-\frac{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}{{\left|h_{\mathrm{ii}}\right|}^2}$

After obtaining iterative formula, minimax transmitting power and precision parameter are set in base station first, according to iterative formula, carry out iterative computation, when this difference that calculates the utility function value that utility function value and last computation obtain is less than the precision parameter of setting, iteration finishes.

For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with the simulation scenarios of view and embodiment, the present invention is described in further detail.

Referring to Fig. 1, introduce the method that the power based on non-cooperative game opinion is controlled.Concrete method flow is:

Step 1, terminal are to base station reporting information, and the information reporting comprises the position of D2D terminal, the distance of D2D terminal and base station, the Initial Trans of D2D terminal;

Calculation of parameter is carried out in step 2, base station, and calculate path gain, user's SINR base station, with this, determines utility function;

Interative computation is carried out in step 3, base station, and optimizing power is controlled;

Power control instruction is assigned to terminal in step 4, base station.

Referring to Fig. 2, introduce the structure of apparatus of the present invention, apparatus of the present invention comprise three main modular:

System information acquisition module, is mainly used in receiving the information of reporting of user base station, upgrades D2D system phone user and D2D user's information;

Data processing module, for calculating path gain between user and user's SINR value, determines utility function;

Power regulation module, for carrying out interative computation according to the result of data processing, constantly adjusts user power and makes it reach Nash Equilibrium point, and power is adjusted to result is issued to each user;

Referring to Fig. 3, introduce research scene of the present invention: in a cellular network, comprise a phone user and multipair D2D user couple, the multiplexing cellular network ascending resource of D2D user.In simulating scenes, the present invention has used 8 couples of D2D users couple.Meanwhile, table 1 has provided the parameter setting of experiment simulation of the present invention.

The setting of table 1 simulation parameter

Referring to Fig. 4, the distribution map that Fig. 4 is cell terminal.Wherein, base station is positioned at center, i.e. (0,0); Asterisk represents phone user; Green circle and red circle represent respectively D2D user's receiving terminal and transmitting terminal.

For positional information and the initial gain value of terminal, carry out the interative computation of following steps:

1) known power iterative formula, will draw intermediate variable α in Initial Trans substitution formula ₁, vector power p ₁, calculate total utility function U ₁;

2) by p ₁in substitution formula, draw intermediate variable α ₂, vector power p ₂;

3) calculate total utility function U ₂, p ₂give p ₁, calculate U ₂with U ₁difference, if difference is less than precision parameter, finishing iteration, otherwise return to previous step.

Referring to Fig. 5, Fig. 5 implements in power control process, the variation diagram of system spectrum utilance.Along with increasing of iterations, the availability of frequency spectrum is in rising trend, and finally converges to a bit.And the whole availability of frequency spectrum is with not carrying out comparing the lifting that has 7% before power control.

Referring to table 2, table 2 has provided the final optimum results of power control scheme under different Initial Transs.As can be seen from Table 2, no matter initial value is how many, power finally all can be tending towards same point.

Final power in the different initial power situations of table 2

Referring to Fig. 6 and Fig. 7, what Fig. 6 and Fig. 7 showed is the impact of cost function C on the availability of frequency spectrum.Can obtain C value larger, it is fewer that phone user's power system capacity reduces.But along with C value increases, the degradation of whole system.

Claims (5)

1. a Poewr control method that is applied to D2D system, is characterized in that, comprising:

Terminal is to base station reporting information, and the information reporting comprises the position of D2D terminal, the distance of D2D terminal and base station, the Initial Trans of D2D terminal;

Calculation of parameter is carried out in base station, and calculate path gain, user's SINR base station, with this, determines utility function;

Based on non-cooperative game model, by interative computation, obtain power and adjust scheme;

Power control instruction is assigned to terminal in base station, realizes power and controls.

2. utility function as claimed in claim 1, comprises utility function and the overall utility function of unique user, and wherein the utility function of unique user is shown below:

u _i＝B _iln(1+γ _i)-C _ip _i

In formula, B _irepresent system bandwidth, C _ip _ithe cost function that represents increased power, γ _ithe SINR that represents D2D user i, γ _ican represent with following formula, wherein | h _ii| ², | h _ij| ²represent path gain:

${\mathrm{\&gamma;}}_i=\frac{P_i}{P_I+N_0}=\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}$

Owing to considering the effectiveness of whole system, so the design of overall utility function is shown below:

$U=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{u}_i=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}[B_{i}1n(1+{\mathrm{\&gamma;}}_i)-C_i{p}_i].$

3. overall utility function as claimed in claim 2, can also represent with following formula, wherein Γ _i(p) can be considered as the optimal compensation item to D2D user i:

$U=u_i\left(p\right)+\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)=u_i\left(p\right)+{\mathrm{\&Gamma;}}_i\left(p\right)$

First by the optimal compensation to independent variable p _iask local derviation

$\frac{{\&PartialD;\mathrm{\&Gamma;}}_i\left(p\right)}{{\&PartialD;p}_i}=\frac{\&PartialD;\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)}{{\&PartialD;p}_i}$

$=\frac{\&PartialD;\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}[1n(1+{\mathrm{\&gamma;}}_j)-C_j{p}_j]}{{\&PartialD;p}_i}$

$=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{1}{1+{\mathrm{\&gamma;}}_j}\&CenterDot;\frac{{\&PartialD;\mathrm{\&gamma;}}_j}{{\&PartialD;p}_i}$

$=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_j}{\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2}}\&CenterDot;\frac{-{\left|h_{\mathrm{jj}}\right|}^2{\left|h_{\mathrm{ij}}\right|}^2{p}_j}{{(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)}^2}$

$\begin{array}{c}=\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}\frac{-{\left|h_{\mathrm{jj}}\right|}^2{\left|h_{\mathrm{ij}}\right|}^2{p}_j}{{(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)}^2+{\left|h_{\mathrm{jj}}\right|}^2{p}_j(\underset{m\&NotEqual;j}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{mj}}\right|}^2{p}_m+{\mathrm{\&sigma;}}^2)}\\ ={\mathrm{\&alpha;}}_i\end{array}$

In order to use the mode of iteration, in above formula, by the power division that once iteration produces, replace current power division, so above formula has just become a constant for current iteration, its integration becomes:

α _ip _i＝Γ _i(p)

It is as follows that we obtain new utility function

L＝u _i(p)+Γ _i(p)

＝ln(1+γ _i)-Cp _i+α _ip _i

Total utility function under this form meets the non-cooperative game model of the optimal compensation; For realizing the optimal compensation, we adopt the method for iteration to carry out, and constantly adjust each user's transmitting power, and when utility function reaches the Nash Equilibrium point of this non-cooperative game model, total utility function reaches optimum value.

4. interative computation as claimed in claim 1, carries out in the following manner:

The utility function that claim 3 is obtained is to p _iask local derviation, have:

$\frac{\&PartialD;u}{{\&PartialD;p}_i}=\frac{\&PartialD;[1n(1+{\mathrm{\&gamma;}}_i)-C_i{p}_i+\underset{j\&NotEqual;i,j\&Element;\mathrm{\&Omega;}}{\mathrm{\&Sigma;}}{u}_j\left(p\right)]}{{\&PartialD;p}_i}$

$=\frac{1}{1+{\mathrm{\&gamma;}}_i}\frac{{\&PartialD;\mathrm{\&gamma;}}_i}{{\&PartialD;p}_i}-C_i+{\mathrm{\&alpha;}}_i$

$=\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}}\frac{{\left|h_{\mathrm{ii}}\right|}^2}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}|{h_{\mathrm{ji}}|}^2{p}_j+{\mathrm{\&sigma;}}^2}-C_i+{\mathrm{\&alpha;}}_i$

The optimized condition of this iteration is that above formula equals 0, and total utility function is to p _ilocal derviation equals 0, is shown below:

$\frac{1}{1+\frac{{\left|h_{\mathrm{ii}}\right|}^2{p}_i}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}}\frac{{\left|h_{\mathrm{ii}}\right|}^2}{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}+{\mathrm{\&alpha;}}_i-C_i=0$

Can obtain p thus _iiterative formula

$p_i=\frac{1}{C_i-{\mathrm{\&alpha;}}_i}-\frac{\underset{j\&NotEqual;i}{\mathrm{\&Sigma;}}{\left|h_{\mathrm{ji}}\right|}^2{p}_j+{\mathrm{\&sigma;}}^2}{{\left|h_{\mathrm{ii}}\right|}^2}$

After obtaining iterative formula, minimax transmitting power and precision parameter are set in base station first, according to iterative formula, carry out iterative computation, when this difference that calculates the utility function value that utility function value and last computation obtain is less than the precision parameter of setting, iteration finishes.

5. the output control device in D2D system, is characterized in that, comprising:

System information acquisition module, in accordance with the method for claim 1 for upgrading D2D system phone user and D2D user's information;

Data processing module, for calculating path gain between user and user's SINR value, determines utility function in accordance with the method for claim 1;

Power regulation module, adjusts scheme for carry out interative computation according to the result of data processing module to determine user power in accordance with the method for claim 4, and power is adjusted to result is issued to each user.

Images